Google Flu Trends, which launched in the US last year, resulted from our observation that certain flu-related search queries are common during flu season, and that searches for flu-related topics are closely correlated to the actual spread of flu. If we tally each day's flu-related search queries, we can estimate how many people have a flu-like illness.
While some traditional flu tracking systems may take days or weeks to collect and release data, Google search queries can be counted immediately. As a result, Google Flu Trends can be updated daily and may provide early detection of flu outbreaks.
To create Google Flu Trends for Australia, we worked with historical seasonal flu data from the Victorian Infectious Diseases Reference Laboratory (VIDRL), as well as our own anonymised, aggregated historical search data, to build a flu model for the state of Victoria. We then extrapolated this model to produce flu models at a national and state level for the rest of Australia, apart from Tasmania and Northern Territory for which we don't have a large enough volume of search queries to be accurate.
Upon building this model, we found that our historical Google search data correlated well with the VIDRL's historical data in terms of flu activity in past years. As the season unfolds this year, we'll be watching to see how well our flu model performs compared with these existing surveillance systems.
Influenza is thought to generate hundreds of thousands of GP visits in Australia every year, and thousands of serious illnesses and sometimes even deaths. Early detection is critical to helping health officials respond more quickly and save lives.
We hope this information complements existing flu surveillance efforts and proves useful for individuals and public health organisations
No comments:
Post a Comment